Simon Inger and Tracy Gardner released an interesting report a little while ago on How scholars navigate to scholarly content. This is a followup to a similar study carried out in 2005 [pdf], and one of the interesting strands of this report is an account of changes in that period.

The focus is on how publishers should think about their network presence in light of changing network behaviors of scholars. They report that readers are increasingly more likely to land in a publisher's website from some other starting point (RSS, Google, A&I database, library portal, etc). This switches focus from navigation of the publisher website to effective disclosure (my word) to those other starting points. They suggest that the "most highly sought-after features of journal web sites are content alerting services, but not personalization and not search functions". They emphasizes the importance of link and data syndication strategies to increase the exposure of their content to their potential readers.

There is much of interest in the specific results, and they have been collected into a readable and brief report. The conclusion provides a good summary.

A key measure of publisher success is the usage of its e-journals, which can be maximised by influencing and enabling all the routes to its content. Library technology plays a key role in user navigation, as well as the more apparent starting points such as Google or major subject A&I databases.
Publishers need to support all conceivable routes to their content through the web. This can best be achieved through the open distribution of XML metadata catalogues, through RSS feeds, collaboration with CrossRef, library technology vendors and through working with major gateways, A&Is and search engines.
Just as was stated in 2005, as metadata distribution is maximised and users are able to choose more freely their preferred routes to content, many of the advanced features that users require are likely to migrate to their chosen gateways (or portals) leaving the publisher site ever more as a content silo, without the need for many of the advanced features that are currently present there.
At the same time it remains true that publishers are under pressure from editorial boards, society members and perversely even from librarians, to create a high level of functionality and the publisher has to manage a careful balancing act to satisfy all of its constituencies. [How readers navigate to scholarly content PDF]

One question I had as I was reading it. They make a distinction between A&I services and library web pages as starting points. When the former was made available through the latter, it was not clear to me which way it was counted.

Some takeaways for me:

  • The report provides good news for libraries, especially in relation to the important 'channeling' role of link resolvers. The authors report that nearly 60% of respondents were guided to e-journals by the library over 95% of the time. They note that this is an 'amazing result'.
  • Disclosure to user workflows has been a recurrent theme of this blog, and I was interested in how this was a major theme of the report. Increasingly we have to build services around user workflow, rather than expect them to build their workflow around services.
  • I recognized the truth of the last paragraph in the conclusions above, and smiled at the expanded version in the body of the text where it was noted that features sometimes had to be incorporated to support a 'political position with respect to societies and powerful editorial board members'.

Related entries:

Comments: 4

Oct 01, 2008
Tony Hirst

'the "most highly sought-after features of journal web sites are content alerting services, but not personalization and not search functions"'

So I'm wondering - will things like the omnibox search on Google Chrome, which makes it almost transparent to add a hook into a site search from your browser address bar ( http://www.google.com/support/chrome/bin/answer.py?answer=95653 ), mean that people come to search sites from their browser address bar/omnibox equivalent?

Thinking about my own browser use, I use smart keywords ( http://www.mozilla.org/products/firefox/smart-keywords.html ) all the time, but not any of the search engines (other than the Google default) that are available from the browser search box...

Oct 01, 2008
Jenny Reiswig

This was a great report - thank you for calling attention to it.

I would expect that "general web search engines" (aka Google) will continue to rise in popularity. Currently, Google searches for scholarly materials generally lead first to the version of the content on the publisher's site and only rarely, if ever, link to that content inside full-text databases provided by 3rd parties. So... in order to simplify link resolving (and save money) we go with aggregators, which then causes problems for Google users. Hmm.

Oct 02, 2008
Michael Jubb

I too read this report with interest. And yes, it does give some good news stories for libraries. But I also recall a presentation that Mayur Amin, of Elsevier, gave to the 2nd Bloomsbury Conference on e-publishing in June this year. He talked among other things about how readers get to Science Direct. What was interesting, as I recall, is that while navigation to journal home pages was overwhelmingly via OPACs and library portals, in navigation to journal articles, libraries did not feature prominently at all. Perhaps this is not surprising, but it does perhaps also send a warning to libraries.


The RIN has commissioned some work on usage of journals and articles in different subjects and universities in the UK, using log analysis; and one of the issues we are looking at is how readers get to the journals and articles they read. We should have some findings available in the New Year.

Oct 03, 2008
Pat Sanner

I'm curious . . . if you were to mash the results of this report with the Ithaca report on faculty reliance (or growing lack of reliance) on libraries and the 2007 OCLC report on growing satisfaction of college students with search engines (even over library reference resources), how would you interpret the interconnectedness of the results?

Thanks